The web site of National Diet Library in Japan provides a lot of early-modern (AD1868-1945) Japanese printed books to the public, but full-text search is essentially impossible. In order to perform advanced search for historical literatures, the automatic textualization of the images is required. However, the ruby system, which is peculiar to Japanese books, gives a serious obstacle against the textualization. When we apply existing OCRs to early-modern Japanese printed books, the recognition rate is extremely low. To solve this problem, we have already proposed a multi-font Kanji character recognition method using the PDC feature and an SVM. In this paper, we propose a ruby character removal method for early-modern Japanese printed books using genetic programming, and evaluate our multi-fonts Kanji character recognition method with 1,000 types of early-modern printed Kanji characters.

Sparse matrix-vector multiplication on GPUs faces to a serious problem when the vector length is too large to be stored in GPU's device memory. To solve this problem, we propose a novel software-hardware hybrid method for a heterogeneous system with GPUs and functional memory modules connected by PCI express. The functional memory contains huge capacity of memory and provides scatter/gather operations. We perform some preliminary evaluation for the proposed method with using a sparse matrix benchmark collection. We observe that the proposed method for a GPU with converting indirect references to direct references without exhausting GPU's cache memory achieves 4.1 times speedup compared with conventional methods. The proposed method intrinsically has high scalability of the number of GPUs because intercommunication among GPUs is completely eliminated. Therefore we estimate the performance of our proposed method would be expressed as the single GPU execution performance, which may be suppressed by the burst-transfer bandwidth of PCI express, multiplied with the number of GPUs.

Cover song detection is becoming a very hot research topic when plentiful personal music recordings or performance are released on the Internet. A nice cover song recognizer helps us group and detect cover songs to improve the searching experience. The traditional detection is to match two musical audio sequences by exhaustive pairwise comparisons. Different from the existing work, our aim is to generate a group of concatenated feature sets based on regression modeling and arrange them by indexing-based approximate techniques to avoid complicated audio sequence comparisons. We mainly focus on using exact locality sensitive mapping (ELSM) to join the concatenated feature sets and soft hash values. Similarity-invariance among audio sequence comparison is applied to define an optimal combination of several audio features. Soft hash values are pre-calculated to help locate searching range more accurately. Furthermore, we implement our algorithms in analyzing the real audio cover songs and grouping and detecting a batch of relevant cover songs embedded in large audio datasets.

We develop a content-based audio COver Song IdeNtification (COSIN) system to detect/group cover songs.The COSIN takes music audio content as input and performs similarity searching to locate variants of the input (i.e., cover versions). Identified cover songs are returned in the rank order according to their similarity to the input.The COSIN also incorporates a set of tools to evaluate retrieval performance so researchers can explore different retrieval schemes and parameters (e.g. recall, precision).The COSIN utilizes a suite of techniques to detect cover songs including: Pitch + Dynamic Programming (DP), Chroma + DP, and Semantic Feature Summarization (SFS) + Hash-Based Approximate Matching (HBAM). Demonstration system shows that COSIN is a very potential music content retrieval tool. Running some music retrieval schemes on COSIN platform, recent experiments with SFS + LSH Variants demonstrate a nicely balanced efficiency (search speed) v. performance (search accuracy) tradeoff.

Visualization plays an important role in PSEs. Some PSE for the support of scientific simulation provides visualization I/F for simulation results. Without deep knowledge of visualization presentation, users require automatic viewpoint selection of the resultant visualization of simulation results. In this paper, we propose an automatic viewpoint selection method, called viewpoint potential, and show some experimental results.

Applying conventional compiler design methodology to a parallelizing compiler, its internal data structures tend to get too complicated quickly. Thus, we introduce object oriented design from the problem analysis stage and achieve the following improvements: 1) consistent modeling from the theory to the implementation; 2) reduced program size through improved reusability based on better class design methodology; and 3) flexible coding through a stub class for parallelization. We extract objects based on stream and thus clarify similarities and differences to the conventional design schemes. Then we show the framework of the internal classes used in our parallelizing compiler which enhances design level reusability in C++ implementation of the compiler. Furthermore, we provide a virtual class to be used as a unit of MIMD style parallel execution and make it a common representation among different parallelization algorithms. Finally, we compare our design against SUIF, which also uses C++, and show the improvements on the design classes.

ASURA is a large scale, chwier-based, distributed, shared memory, multiprocessor being developed at Kyoto University and Kubota Corporation.Up to 128 clusters are interconnected to form an ASURA system of up to 102d processors.The basic concept of the ASURA design is to take advantage of the hierarchical structure of the system.Implementing this concept, a large shared cache is piaced between each cluster and the inter-ciuster network.The shared cache and the shared memories distributed among the clusters form part of ASURA's hierarchical memory architecture, providing various unique features to ASURA.In this paper, the hierarchicalmemory architecture of ASURA and its unique cache coherence scheme, including a proposal of a new hierarchical directory scheme, are described wzth some simulation results.Permission tocopy without fes all or pan of Ibis .n81e&al is granted, pmvidd that tbe copies me not made or dishiluted fw dim commenial advantage, the ACMwpyrisbtmalice and tbe title of the F.Iblication and its date appesr, snd notice is given that cqying is by permission of tbe Association forComputing Mschimy.ToCOPY ciherwi%csto republih.requires a fee Snlvcu Specilic pmnissian.

Recurrent neural networks with arbitrary feedback connections are highly nonlinear dynamical systems exhibiting variegated complex dynamical behavior. The applications of this temporal behavior hold possibilities for information processing. Supervised learning for recurrent networks is studied with emphasis on learning aperiodic motions. APOLONN (adaptive nonlinear pair oscillators with local connections) is used for speech synthesis. The naturalness of a human's voice seems to come from fluctuations in voice source waveforms. The authors trained APOLONN to learn the voice source waveforms, including fluctuations of amplitudes and periodicities. After the learning, APOLONN was able to generate the waveforms with fluctuations. APOLONN can also generate waveforms with modulated amplitudes and frequencies by a simple scaling of the parameters. The results encourage further applications of recurrent networks

Abstract This paper describes how new learning methods may make it possible for a large‐scale, hierarchical neural network to recognize most Japanese handwritten characters. This is a very large and complex task, as the Japanese character set consists of about 3000 categories which can be written in many different ways. Such a difficult task can lead a neural network to converge very slowly and to yield recognition rates that are uneven between categories. To address these problems we here propose five learning methods as modifications of the conventional back‐propagation learning rule. These methods produce fast convergence, even recognition rates over all categories, and adequate recognition of test samples. We also describe how a large‐scale neural network can be built by dividing the recognition task into several subtasks, with networks for each subtask, and then integrating these subnetworks in a large network with a hierarchical structure. In a hierarchical network, the upper level network directly integrates outputs from each lower level network. Application of that network to handwritten Japanese character recognition has resulted in poor recognition, because lower level networks do not know about unknown input patterns, and the direct integration of ambiguous outputs from many lower level networks confuses the upper level network. We propose a new integration method which provides each subnetwork with more information as to how close an input pattern is to the categories of that subnetwork. This method resulted in high recognition performance for character recognition. We here described the above methods, and report the performance of our implementation of a neural network for the recognition of 71 Hiragana characters, and describe our implementation of this network on a hypercube concurrent computer.